It never ceases to amaze me the depth of the world economy, and how seemingly small niches can add up to billion dollar markets.
But $500mm is a tremendous amount of money to raise pre-revenue, which would really only be appropriate in an area with tremendous barriers to entry. Gaming physics, AI, map generation, etc. all seem like exactly the kind of thing that you can break apart into smaller pieces to get to market quickly and iteratively.
It's not like a molecule that needs to through Phase 3 trials, or a rocket which needs to make orbit, or something with a huge CapEx component. Even rockets can scale up from computer models, to scale models, to rough prototypes, etc.
I wouldn't go as far as to say the Improbable deal made sense, but your analysis here is pretty shallow.
SpatialOS isn't really something that can be built as a series of libraries a la carte and individually proven. It's an opinionated and tightly integrated approach to solving a very difficult problem, and the entire point is the turnkey platform. Either it's ready to build notable games on, or it's not.
Further, Improbable wants developers to pull the trigger on spending millions or even tens of millions developing titles for the platform. Investing this much in a platform that fails is an existential crisis for a games company. There's no way your half built SpatialOS game can run anywhere else.
Therefore, confidence is key, and raising an eye watering amount of money is basically table stakes to be taken seriously.
Whether it's worth solving a problem that can't be broken down in viable independent chunks is another matter (I have no interest in that personally). I'm not saying I'd invest at that valuation either. But it's certainly not as straight forward as you seem to be claiming.
Actually I saw a presentation about pharmacy companies already aquiring startups at Phase 1 or even after pre-clinical trials, so VC money isn't needed post Phase 1 generally.
IBM didn't see MS coming, MS didn't see Google coming, Google didn't see FB coming, etc.
The beauty of this is that things look done and then someone see a new niche and it turns out it wasn't a niche, but something millions of people wanted.
But $500mm is a tremendous amount of money to raise pre-revenue, which would really only be appropriate in an area with tremendous barriers to entry. Gaming physics, AI, map generation, etc. all seem like exactly the kind of thing that you can break apart into smaller pieces to get to market quickly and iteratively.
It's not like a molecule that needs to through Phase 3 trials, or a rocket which needs to make orbit, or something with a huge CapEx component. Even rockets can scale up from computer models, to scale models, to rough prototypes, etc.